centralize -Zmin-function-alignment logic
#142854
Merged
+10
−17
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
rustbot
added
A-attributes
|
Some changes occurred to the CTFE machinery Some changes occurred to the CTFE / Miri interpreter cc @rust-lang/miri Some changes occurred in compiler/rustc_codegen_ssa Some changes occurred in compiler/rustc_codegen_ssa/src/codegen_attrs.rs |
|
r? codegen |
98 tasks
|
This looks good to me. r? workingjubilee |
bors
added
S-waiting-on-bors
jhpratt
added a commit
to jhpratt/rust
that referenced
this pull request
Jun 22, 2025
…-alignment, r=workingjubilee centralize `-Zmin-function-alignment` logic tracking issue: rust-lang#82232 discussed in: rust-lang#142824 (comment) Apply the `-Zmin-function-alignment` value to the alignment field of the function attributes when those are created, so that individual backends don't need to consider it. The one exception right now is cranelift, because it can't yet set the alignment for individual functions, but it can (and does) set the global minimum function alignment. cc `@RalfJung` I think this is an improvement regardless, is there anything else that should be done for miri?
bors
added a commit
that referenced
this pull request
Jun 23, 2025
Rollup of 6 pull requests Successful merges: - #140136 (Add an aarch64-msvc build running on ARM64 Windows) - #141597 (Document subdirectories of UI tests with README files) - #142823 (Port `#[no_mangle]` to new attribute parsing infrastructure) - #142828 (1.88.0 release notes) - #142854 (centralize `-Zmin-function-alignment` logic) - #142875 (Check rustdoc-json-types FORMAT_VERSION is correctly updated) r? `@ghost` `@rustbot` modify labels: rollup
bors
added a commit
that referenced
this pull request
Jun 23, 2025
Rollup of 5 pull requests Successful merges: - #141597 (Document subdirectories of UI tests with README files) - #142823 (Port `#[no_mangle]` to new attribute parsing infrastructure) - #142828 (1.88.0 release notes) - #142854 (centralize `-Zmin-function-alignment` logic) - #142875 (Check rustdoc-json-types FORMAT_VERSION is correctly updated) r? `@ghost` `@rustbot` modify labels: rollup
rust-timer
added a commit
that referenced
this pull request
Jun 23, 2025
Rollup merge of #142854 - folkertdev:centralize-min-function-alignment, r=workingjubilee centralize `-Zmin-function-alignment` logic tracking issue: #82232 discussed in: #142824 (comment) Apply the `-Zmin-function-alignment` value to the alignment field of the function attributes when those are created, so that individual backends don't need to consider it. The one exception right now is cranelift, because it can't yet set the alignment for individual functions, but it can (and does) set the global minimum function alignment. cc ``@RalfJung`` I think this is an improvement regardless, is there anything else that should be done for miri?
workingjubilee
added a commit
to workingjubilee/rustc
that referenced
this pull request
Jun 23, 2025
…no-attributes, r=workingjubilee fix `-Zmin-function-alignment` on functions without attributes tracking issue: rust-lang#82232 related: rust-lang#142854 The minimum function alignment was skipped on functions without attributes (because the logic was in a loop that only runs if there is at least one attribute). The underlying reason we didn't catch this before is that in our testing we generally apply `#[no_mangle]` to functions that are tested. I've added a test now that deliberately has no attributes. r? `@workingjubilee`
rust-timer
added a commit
that referenced
this pull request
Jun 24, 2025
Rollup merge of #142923 - folkertdev:min-function-alignment-no-attributes, r=workingjubilee fix `-Zmin-function-alignment` on functions without attributes tracking issue: #82232 related: #142854 The minimum function alignment was skipped on functions without attributes (because the logic was in a loop that only runs if there is at least one attribute). The underlying reason we didn't catch this before is that in our testing we generally apply `#[no_mangle]` to functions that are tested. I've added a test now that deliberately has no attributes. r? `@workingjubilee`
RalfJung
reviewed
Jun 25, 2025
| @@ -126,6 +126,12 @@ fn codegen_fn_attrs(tcx: TyCtxt<'_>, did: LocalDefId) -> CodegenFnAttrs { | |||
| } | |||
| } | |||
|
|
|||
| // Apply the minimum function alignment here, so that individual backends don't have to. | |||
There was a problem hiding this comment.
It's actually more important than that, if you consider that -C flags can differ between translation units.
This ensures functions defined in a crate compiled with -Cmin-function-alignment get that alignment independent of the crate they are codegen'd in.
There was a problem hiding this comment.
If a function from crate A is generated in another crate B compiled with -Zmin-function-alignment=B, I would expect the alignment to be at least B.
There was a problem hiding this comment.
We don't say anything about which function is generated where, so I don't see how such a guarantee could be practically useful?
From a const-eval perspective, doing this would be just unsound as it means the result of "which alignment does this function have" could be different in different parts of the program.
RalfJung
reviewed
Jun 25, 2025
| // Function alignment can be set globally with the `-Zmin-function-alignment=<n>` flag; | ||
| // the alignment from a `#[repr(align(<n>))]` is used if it specifies a higher alignment. | ||
| let fn_align = self.tcx.codegen_fn_attrs(instance.def_id()).alignment; | ||
| let global_align = self.tcx.sess.opts.unstable_opts.min_function_alignment; |
There was a problem hiding this comment.
Is there any way we can have a lint against accessing tcx.sess.opts in the interpreter? Or is there some way the query machinery could help us ensure this? This is a soundness bug waiting to happen. :/ And it can't be just the interpreter, right? Other queries also must be cross-crate-stable in the same way?
This was referenced Jun 25, 2025
samueltardieu
added a commit
to samueltardieu/rust
that referenced
this pull request
Jun 26, 2025
codegen_fn_attrs: make comment more precise Follow-up to rust-lang#142854. r? `@oli-obk` or `@workingjubilee`
github-actions bot
pushed a commit
to rust-lang/rustc-dev-guide
that referenced
this pull request
Jun 26, 2025
Rollup of 5 pull requests Successful merges: - rust-lang/rust#141597 (Document subdirectories of UI tests with README files) - rust-lang/rust#142823 (Port `#[no_mangle]` to new attribute parsing infrastructure) - rust-lang/rust#142828 (1.88.0 release notes) - rust-lang/rust#142854 (centralize `-Zmin-function-alignment` logic) - rust-lang/rust#142875 (Check rustdoc-json-types FORMAT_VERSION is correctly updated) r? `@ghost` `@rustbot` modify labels: rollup
workingjubilee
added a commit
to workingjubilee/rustc
that referenced
this pull request
Jun 26, 2025
codegen_fn_attrs: make comment more precise Follow-up to rust-lang#142854. r? ``@oli-obk`` or ``@workingjubilee``
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
tracking issue: #82232
discussed in: #142824 (comment)
Apply the
-Zmin-function-alignmentvalue to the alignment field of the function attributes when those are created, so that individual backends don't need to consider it.The one exception right now is cranelift, because it can't yet set the alignment for individual functions, but it can (and does) set the global minimum function alignment.
cc @RalfJung I think this is an improvement regardless, is there anything else that should be done for miri?